Cyclophosphamide enhances antitumor efficacy of oncolytic adenovirus expressing uracil phosphoribosyltransferase (UPRT) in immunocompetent Syrian hamsters

Authors


Correspondence to: Masato Abei, Division of Gastroenterology, Faculty of Medicine, University of Tsukuba, Tennoudai 1-1-1, Tsukuba-shi, Ibaraki 305–8575, Japan. Fax: +81–298-53–3128, E-mail: m-abei@md.tsukuba.ac.jp

Abstract

Oncolytic viruses (OVs) are novel cancer therapeutics with great promise, but host antiviral immunity represents the hurdle for their efficacy. Immunosuppression by cyclophosphamide (CP) has thus been shown to enhance the oncolytic efficacy of many OVs, but its effects on OVs armed with therapeutic genes remain unknown. We have previously reported on the efficacy of AxE1CAUP, an oncolytic adenovirus (OAd) expressing uracil phosphoribosyltransferase (UPRT), an enzyme that markedly enhanced the toxicity of 5-fluorouracil (5-FU), in immunodeficient, Ad-nonpermissive nude mice. Here we explored the efficacy and safety of intratumoral (i.t.) AxE1CAUP/5-FU therapy and of its combination with CP for syngenic HaP-T1 pancreatic cancers in immunocompetent, Ad-permissive Syrian hamsters. AxE1CAUP infected, replicated, expressed UPRT, and increased the sensitivity to 5-FU in HaP-T1 cells in vitro. I.t. AxE1CAUP/5-FU treatment inhibited the growth of subcutaneous HaP-T1 allografts. The combination with high-dose CP inhibited serum Ad-neutralizing antibody formation, increased intratumoral AxE1CAUP replication and UPRT expression, and resulted in further enhanced therapeutic effects with 5-FU. Neither body weight nor histology of the liver and lung changed during these treatments. A clinically-approved, intermediate-dose CP also enhanced the efficacy of i.t. AxE1CAUP/5-FU treatment in these hamsters, which was not affected by preexisting immunity to the vector. These data demonstrate the excellent antitumor efficacy and safety of an OAd armed with a suicide gene in combination with CP for treating syngenic tumors in immunocompetent, Ad-permissive animals, indicating the efficacy of CP in overcoming the hurdle of antiviral immunity for effective OV-mediated gene therapy.

Abbreviations
Ad

adenovirus

Ad5-wt

wild-type adenovirus type 5

CAR

coxsackievirus adenovirus receptor

CP

cyclophosphamide

5-FU

5-fluorouracil

HSVtk

herpes simplex virus thymidine kinase

i.m.

intramuscular

i.p.

intraperitoneal

i.t.

intratumoral

MOI

multiplicity of infection

OAd

oncolytic adenovirus

OV

oncolytic virus

PFU

plaque forming unit(s)

s.c.

subcutaneous

UPRT

uracil phosphoribosyltransferase

Oncolytic viruses (OVs), which selectively replicate in and lyse tumor cells, are a novel class of anticancer therapeutics with great promise.[1-4] The development of an effective combination therapy with OVs to further enhance their antitumor effects has been a recent topic.[5] OVs armed with suicide genes (or prodrug converting-enzyme genes),[6-10] cytokine genes,[11-13] or tumor suppressor genes[14] are expected to enhance the antitumor efficacy further.

We have previously reported the preclinical efficacy of several oncolytic adenoviruses (OAds) and their gene therapy applications.[9, 15-17] In particular, we and others have explored the efficacy of AxE1CAUP, an E1B-55 kDa-defective OAd expressing uracil phosphoribosyltransferase (UPRT; EC 2.4.2.9), which directly converts 5-fluorouracil (5-FU) into 5-fluorouridine monophosphate and greatly enhances the cytotoxicity of 5-FU.[9, 10] Although 5-FU is a widely used chemotherapeutic agent, it requires several steps of enzymatic conversion to its active metabolites, FUTP and FdUMP, in cancer cells to exert its cytotoxic activity, which underlies the tolerance of cancers to 5-FU.[18] We and others have found that the transduction of UPRT markedly sensitizes cancer cells to a low concentration of 5-FU[9, 10] and prolongs the overall survival of nude mice bearing human cancer xenograft,[9] suggesting the effectiveness of this approach in overcoming the resistance to 5-FU. In addition, the timing of 5-FU administration significantly affected the efficacy of the therapy.[9] However, how OAds might behave in humans cannot be estimated from experiments in immunodeficient, Ad-nonpermissive animals, such as nude mice, and should be tested using an immunocompetent, Ad-permissive animals such as Syrian hamsters[19, 20] in which the effects on both antitumor and antiviral immunity can also be addressed.

On the other hand, the development of adaptive antiviral immunity represents the major hurdle for the delivery and spread of OVs in tumors with neutralizing antibodies (Nab) and cytotoxic T cells blocking vascular delivery and intratumoral (i.t.) spread.[21-23] Thus, the effects of serum Nab on the efficacy and safety of OAds have been discussed in immunodeficient[24, 25] and immunocompetent[26, 27] models. Cyclophosphamide (CP) is an alkylating agent that is commonly used as an anticancer chemotherapeutic agent and as an immunosuppressive agent. Previous studies have demonstrated that CP has different effects on the immune system according to the dose administered. It has been reported that low-dose CP enhances antitumor immune responses via selective depletion of regulatory T cells that suppress antitumor CD8+ T-cell responses.[28-30] In striking contrast, high-dose CP causes immunosuppression by ablating cytotoxic T-cell and helper T-cell functions and preventing the Nab formation. CP has been shown to enhance the oncolytic efficacy of several OVs including herpes simplex virus,[31, 32] Ad,[33] reovirus,[34] measles[35] and vaccinia virus.[36] However, how immunosuppression by CP influences the overall efficacy of the OAd armed with therapeutic genes remains totally unknown.

Pancreatic cancer is ranked as the fourth leading cause of cancer-related death in the United States.[37] Since early diagnosis of pancreatic cancer is difficult, only 9% of the patients undergo curative surgery[38] and current chemotherapeutic regimens have limited efficacy leading to their high mortality rates. A new mode of treatment, such as OV therapy, is clearly necessary to improve the survival of these patients.

In the present study, we aim to evaluate the antitumor efficacy and safety of i.t. AxE1CAUP/5-FU treatment in a syngenic pancreatic cancer model in immunocompetent, Ad-permissive Syrian hamsters. In addition, we examined the effects of a high-dose and a clinically-approved intermediate-dose of CP on the efficacy of the i.t. AxE1CAUP/5-FU treatment in hamsters with or without preexisting immunity to the vector. The results support the efficacy of CP in overcoming the hurdle of antiviral immunity for successful OV-mediated gene therapy for cancers.

Material and Methods

Cell lines, culture, and adenoviral vectors

The Syrian hamster pancreatic cancer cell line, HaP-T1, and the human pancreatic cancer lines, PANC-1 and MIA Paca2, were purchased from RIKEN Cell Bank (Tsukuba, Ibaraki, Japan) and cultured under recommended conditions. AxCAZ3, a recombinant Ad containing a LacZ gene,[39] and AxE1CAUP were constructed as described previously.[9, 10] Quantification of viruses was performed using an Adeno-X rapid titer kit (Clontech, Palo Alto, CA).

Western blotting of coxsackievirus adenovirus receptor (CAR) and UPRT

Western blotting of CAR and UPRT was performed as described previously.[9, 17, 40] Cell lysates were prepared as described previously.[17] Tissue samples were homogenized in lysis buffer containing 75 mM HEPES (pH 7.4), 1.5 mM EGTA, 150 mM KCl, 1.5 mM MgCl2, 15% glycerol, 0.05% NP-40 and were centrifuged at 10,000g for 10 min and the supernatant was used for Western blotting.

Ad-mediated gene transduction, viral replication, and sensitivity to 5-FU in vitro

To evaluate Ad-mediated gene transduction efficiency, cells in 96-well plates (1 × 104 cells/well) were infected with AxCAZ3, at 1, 10, 100 MOIs. After 24 hr, the cells were lysed, centrifuged and the activity of β-galactosidase in the supernatant was measured using β-gal reporter gene assay (Life Technologies, Carlsbad, CA). To evaluate the replication of AxE1CAUP, cells in 12-well plates (1 × 104 cells/well) were infected with AxE1CAUP at 10 MOI. The cells and culture supernatant were harvested on 2, 4, and 6 days after the infection. Quantitative real-time PCR assays for the Ad E1A DNA were carried out using the Applied Biosystems StepOnePlus instrument (Life Technologies) as described previously.[41] In vitro sensitivity of cancer cells to 5-FU with and without UPRT transduction was evaluated as described previously.[9, 10] Briefly, cells with or without infection with AxE1CAUP (10 MOI) were exposed to various concentrations of 5-FU (0.1–1,000 μM/l) 24 or 96 hr later and incubated at 37°C for another 3 days. Viable cells were quantitated by WST-1 assay. The 50% inhibitory concentration (IC50) of 5-FU was calculated using curve-fitting parameters for each set of conditions.

Immunosuppression by CP

In the first experiment, the initial dose of CP (Sigma-Aldrich) of 140 mg/kg and 100 mg/kg twice weekly thereafter (high-dose CP) was intraperitoneally (i.p.) injected as previously reported.[26, 27, 33] This CP dose, however, was considered higher than clinically-approved regimens of CP.[42] Thus, in the second experiment, we i.p. injected a reduced intermediate-dose of CP (37 mg/kg) twice weekly, which is equivalent to one of the standard human CP protocols (5 mg/kg twice weekly), according to the human-to-animal conversion factors recommended by the US Food and Drug Administration.[43]

Immunization before treatment and titration of serum Ad neutralizing antibody (Nab)

Preexisting immunity was generated by a single intramuscular (i.m.) injection of AxE1CAUP (1 × 109 PFU). Hamsters received subcutaneous injection of HaP-T1 cells (106 cells/site) 14 days after the preimmunization. When tumors reached the volume of approximately 100 mm3 at about 7 days later, the animals received i.t. injection of AxE1CAUP (3 × 108 PFU/tumor × 3 times). Thus, the treatment was started 3 weeks after the pre-immunization. Ad Nab titers were examined as described previously.[26, 44] NAb titer for each serum sample was described as the reciprocal dilution that inhibited AxCAZ3-mediated transduction by 50%. Sera that did not show inhibition of the transduction at the lowest dilution tested (1:16) were considered as undetectable (zero).

Animal studies

Animal studies were approved by the Animal Study Committee of University of Tsukuba and were conducted in the Laboratory Animal Resource Center of the University in accordance with institutional and national regulations. Six-week-old female Syrian hamsters were obtained from SLC Japan (Tokyo, Japan) and were quarantined for 1 week. Hamsters were injected subcutaneously with 106 HaP-T1 cells/site. The first animal study was planned to clarify the efficacy and safety of i.t. AxE1CAUP/5-FU treatment with or without immunosuppression by high-dose CP and included the following seven groups: (i) PBS, (ii) 5-FU, (iii) AxE1CAUP, (iv) AxE1CAUP/5-FU, (v) high-dose CP, (vi) AxE1CAUP/high-dose CP, and (vii) AxE1CAUP/high-dose CP/5-FU groups. To evaluate the systemic antitumor immune responses, subcutaneous HaP-T1 tumors were generated in the bilateral flanks of each hamster. When tumors reached the volume of approximately 100 mm3, hamsters were randomized and one of the tumors in each animal was infected with i.t. AxE1CAUP at a dose of 3 × 108 PFU/tumor/day (total 9 × 108 PFU/tumor) or PBS on days 1 to 3 with or without i.p. 5-FU administration (20 mg/kg/d) on days 7 to 11. Three animals from each group were sacrificed on days 7 and 30 to harvest the sera, the tumors, and the organs. For histological analysis, the tumor, liver and lung were fixed in 4% formalin and 4-Am tissue sections were stained with hematoxylin and eosin. The second animal study was planned to clarify the effects of intermediate dose of CP and the preexisting immunity to Ad5 on the efficacy of i.t. AxE1CAUP/5-FU treatment, and included the following 11 groups: (i) non-immunized (non-imm)/PBS, (ii) pre-immunized (pre-imm)/PBS, (iii) non-imm/intermediate-dose CP, (iv) non-imm/5-FU, (v) non-imm/AxE1CAUP/5-FU, (vi) non-imm/AxE1CAUP/ intermediate-dose CP, (vii) non-imm/AxE1CAUP/intermediate-dose CP/5-FU, (viii) non-imm/AxE1CAUP, (ix) pre-imm/AxE1CAUP/5-FU, (x) pre-imm/AxE1CAUP/intermediate-dose CP, and (xi) pre-imm/AxE1CAUP/intermediate-dose CP/5-FU groups. The same doses and regimens were used for the administrations of i.t. AxE1CAUP and i.p. 5-FU as in the first experiment. The sera were obtained from three animals in each group on day 30.

Immunohistochemistry and quantitation of UPRT mRNA in the tumors

The intratumoral presence of Ad was examined by immunohistochemistry of E1A performed using a rabbit polyclonal antibody against E1A (13S-5; Santa Cruz Biotechnology, Santa Cruz, CA) and a Vectastain ABC-PO (rabbit-IgG) kit (Vector Laboratories, Burlingame, CA). Sections were counterstained with hematoxylin. Total RNA in the tumors was extracted with Trizol Reagent (Life Technologies) and was reverse transcribed into cDNA with Superscript III (Life Technologies). The UPRT mRNA was quantitated by quantitative revere transcription (RT)-PCR assay using an Applied Biosystems StepOnePlus instrument (Life Technologies). β-actin gene was used as an endogenous control and the comparative Ct method was applied. The sequences of specific primers used for UPRT and β-actin were as follows: UPRT sense, 5′-TCTGCGTGCGGGTCTTG-3′ and antisense, 5′-CGCGCTCGGAACGTTTT-3′; β-actin sense, 5′-GGCCGTGCTGTCCCTGTAT-3′ and antisense, 5′-CCGGAGTCCATCACAATGC-3′.

Statistical analysis

The significance of differences between groups was assessed by Student's unpaired two-tailed t-test and was considered statistically significant when adjusted p values were <0.05.

Results

HaP-T1 cells express CAR and show excellent Ad-mediated gene transduction efficiency

At first, we found high expression of CAR in HaP-T1 hamster pancreatic cancer line as well as in two human pancreas cancer lines (PANC-1, MIA Paca2) on Western blots (Fig. 1a). The β-gal assay after infection with AxCAZ3 showed dose-dependent excellent Ad-mediated gene transductions in these three lines with a significantly (p < 0.01) better efficiency in HaP-T1 line than in the two human counterparts (Fig. 1b).

Figure 1.

Expression of CAR, Ad transduction efficiency, replication of AxE1CAUP, AxE1CAUP-induced UPRT expression and enhanced sensitivity to 5-FU in HaP-T1 cells. (a) Expression of CAR in hamster (HaP-T1) and human (PANC-1, MIA Paca2) pancreas cancer lines was examined by Western blotting. β-actin was stained as internal control. (b) Transduction efficiency of Ad was examined by β-galactosidase activity 24 hr after infection with a LacZ-expressing Ad vector, AxCAZ3. β-galactosidase activity was expressed as relative light units (RLU). **p < 0.01. (c) Replication potency of AxE1CAUP vector in hamster (HaP-T1) and human (PANC-1, MIA Paca2) pancreas cancer cells. Copy numbers of Ad E1A were measured by quantitative PCR as described in Material and Methods. Mean ± SD of three independent experiments are shown. (d) Expression of UPRT after infection of AxE1CAUP in HaP-T1 cells was evaluated by Western blotting. β-actin was stained as internal control. (e, f) Sensitivity in vitro to 5-FU of HaP-T1 cells after infection with AxE1CAUP. HaP-T1 cells were treated with AxE1CAUP (▪) at 10 MOI or PBS (♦); 5-FU was supplied to the cells at 24 hr (e) or 96 hr (f) after the treatment, and then cell viability was determined by WST-1 assay. Percentages (mean ± SD) of surviving cells compared with control cells without 5-FU from four individual experiments are shown.

AxE1CAUP replicates, expresses UPRT and enhances sensitivity to 5-FU in HaP-T1 cells in vitro

AxE1CAUP replicated well in HaP-T1 line as well as in the two human counterparts, according to the quantitative PCR of Ad E1A DNA (Fig. 1c). UPRT expression was detected in HaP-T1 cells at 24 hr and was further increased at 48 hr after the infection (Fig. 1d). The UPRT transduction significantly increased the sensitivity of HaP-T1 cells to 5-FU (Figs. 1e and 1f). When 5-FU administration was started early (24 hr) after the infection, IC50 of 5-FU decreased fivefold from the control without the infection (Fig. 1e, p < 0.01 for AxE1CAUP (IC50, 0.64 μM) versus PBS (IC50, 3.29 μM)). When 5-FU administration was started late (96 hr), the IC50 decreased 13-fold from the control (Fig. 1f, p < 0.01 for AxE1CAUP (IC50, 0.39 μM) versus PBS (IC50, 5.27 μM)) and significantly (p < 0.05) stronger sensitizing effect of AxE1CAUP was observed than when 5-FU was started early (24 hr).

Antitumor efficacy of i.t. AxE1CAUP/5-FU and further enhancement by high-dose CP in hamsters

We compared the in vivo antitumor effects of i.t. AxE1CAUP treatment followed by i.p. 5-FU administration with or without immunosuppression by high-dose CP on subcutaneous HaP-T1 allografts in immunocompetent Syrian hamsters. The tumors treated with PBS (control) or i.p. 5-FU alone grew rapidly by about 18-fold, or 14-fold, respectively, of their initial volume, whereas those in the AxE1CAUP group increased by about ninefold, on day 22 (Fig. 2a, p < 0.01 for AxE1CAUP vs. PBS). Conversely, the AxE1CAUP/5-FU group (fourfold on day 22) showed significantly retarded tumor growth compared with the PBS and the AxE1CAUP groups (Fig. 2a, p < 0.01 for AxE1CAUP/5-FU vs. PBS and p < 0.01 for AxE1CAUP/5-FU vs. AxE1CAUP). Moreover, the combination of high-dose CP and i.t. AxE1CAUP/5-FU treatment (AxE1CAUP/CP/5-FU) further suppressed the tumor growth (1.8-fold on day 22) significantly (p < 0.01) compared with the treatment without CP (Fig. 2a). To evaluate whether the immunosuppression enhanced viral oncolysis, we also compared the effects of i.t. AxE1CAUP treatment alone with or without CP administration. The AxE1CAUP/CP group showed significantly (p < 0.05) inhibited tumor growth (sixfold on day 22) compared with the AxE1CAUP alone group (Fig. 2a). CP alone did not have a significant effect. These results suggest that high-dose CP enhances both the oncolytic effects and the suicide gene effects of the i.t. AxE1CAUP treatment.

Figure 2.

Antitumor effects of i.t. AxE1CAUP with or without i.p. 5-FU and high-dose CP for inoculated and non-innoculated subcutaneous HaP-T1 tumors in Syrian hamsters. Two subcutaneous HaP-T1 tumors were established in bilateral franks and AxE1CAUP was inoculated in one of the tumors in each hamster. The effects of i.t. AxE1CAUP with or without i.p. 5-FU and high-dose CP (140 mg/kg initially and 100 mg/kg/twice weekly thereafter) treatments on the growth of the inoculated (a) and noninoculated (b) tumors were examined. (a) The growth of inoculated tumors receiving the following seven different treatments was compared: (i) PBS (♦), (ii) 5-FU (•), (iii) AxE1CAUP (▴), (iv) AxE1CAUP/5-FU (X), (v) CP (▪), (vi) AxE1CAUP/CP (○), and (vii) AxE1CAUP/CP/5-FU (◻) groups. Mean + SD (n = 6) are shown. *p < 0.05, **p < 0.01. (b) The growth of non-inoculated tumors in the hamsters treated with AxE1CAUP (○) or AxE1CAUP/CP (▴) was compared with the control (PBS; ♦). Mean + SD (n = 6) are shown. **p < 0.01.

Effects of i.t. AxE1CAUP with or without high-dose CP for remote tumors

To investigate whether i.t. AxE1CAUP can induce systemic antitumor immunity, subcutaneous HaP-T1 tumors were generated in bilateral flanks of Syrian hamsters and AxE1CAUP was injected into only one tumor. The i.t. AxE1CAUP inoculation caused mild but significant (p < 0.01) growth inhibition of the remote, contralateral, noninoculated tumors compared with the growth of respective tumors in the control hamsters (PBS): the volume of the noninoculated tumors in the AxE1CAUP group had increased by 12-fold, whereas the tumors in the PBS group increased by 17-fold on day 22 (Fig. 2b). Absence of Ad in the remote tumors was confirmed by no immunohistological staining of Ad E1A (results not shown). Immunosupression by CP eliminated this antitumor effect of AxE1CAUP on the remote tumors (Fig. 2b, p > 0.05 for AxE1CAUP/CP vs. PBS).

High-dose CP prevents the formation of serum neutralizing antibody and promotes i.t. AxE1CAUP replication and UPRT expression

During the in vivo study described above, sera were collected from the animals on days 7 and 30. Serum Ad Nab titers were significantly (p < 0.01) elevated on days 7 and 30 in the AxE1CAUP/5-FU group compared with the PBS group, whereas the titers were significantly (p < 0.01) inhibited in the AxE1CAUP/CP/5-FU group compared with the AxE1CAUP/5-FU group (Fig. 3a). Tumor tissues were obtained from three animals in each group on day 7 (just before the 5-FU administration). The Ad presence was evaluated by immunohistochemical staining of E1A. The expression of UPRT mRNA and protein in the tumor lysates was evaluated by quantitative RT-PCR and Western blotting, respectively. The tumors in the AxE1CAUP/CP group showed enhanced immunohistochemical E1A staining (Fig. 3b) and increased mRNA (Fig. 3c; p < 0.01) and protein (Fig. 3d) expressions of UPRT, compared with those in the AxE1CAUP group. These results demonstrate that high-dose CP enhanced the AxE1CAUP replication and its transgene (UPRT) expression, and account for the enhanced antitumor efficacy of the combination therapy (Fig. 2a).

Figure 3.

Serum Ad neutralizing antibody (Nab), intratumoral Ad replication and UPRT expression in Syrian hamsters treated with i.t. AxE1CAUP with or without high-dose CP. Serum Ad Nab formation following i.t. AxE1CAUP treatment with and without high-dose CP. (a) Serum samples were obtained on day 7 and day 30 from three animals in each group and the Nab titers were quantitated as described in Material and Methods. Sera that did not show inhibition of Ad transduction at the lowest dilution tested (1:16) were considered as undetectable. *p < 0.05, **p < 0.01. (b) Tumor tissues treated with i.t. AxE1CAUP with or without high-dose CP were obtained on day 7 (before 5-FU administration) and the AxE1CAUP vector in the tumors was immunohistochemically stained with anti-E1A antibody. Scale bar represents 100 µm. (c) Relative expression of UPRT mRNA in the tumors was measured by quantitative RT-PCR (comparative Ct methods; mean ± SD (n = 3) are shown). **p < 0.01. (d) The expression of UPRT protein in the tumors was examined by Western blotting. β-actin was stained as internal control. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Safety of i.t. AxE1CAUP/5-FU treatment with or without high-dose CP

Regarding the safety of the treatment, body weight increased gradually in all seven groups of animals (PBS, 5-FU, CP, AxE1CAUP, AxE1CAUP/5-FU, AxE1CAUP/CP, and AxE1CAUP/CP/5-FU groups) and did not vary significantly between the groups (Fig. 4a). Two animals in each group were sacrificed on day 30 and the histology of the lung and liver was examined, as infection with wild type Ad in humans often causes liver or lung injury. No histological abnormality was found in the liver and lung from all groups of hamsters (Fig. 4b).

Figure 4.

Body weight and histology of liver and lung in Syrian hamsters treated with i.t. AxE1CAUP/5-FU with or without high-dose CP. (a) Changes of body weight of the hamsters in the following seven groups are shown: (i) PBS (♦), (ii) 5-FU (•), (iii) AxE1CAUP (▴), (iv) AxE1CAUP/5-FU (X), (v) CP (▪), (vi) AxE1CAUP/CP (○), and (vii) AxE1CAUP/CP/5-FU (◻) groups. Mean (n = 6) percentages of body weight as compared with the initial body weight are shown. (b) Histological examination (H.E. stain, original magnification ×40) of the liver and lung of the hamsters in the PBS (control) group, i.t. AxE1CAUP/5-FU group, or i.t. AxE1CAUP/CP/5-FU group. Scale bar represents 100 µm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Clinically approved intermediate-dose CP also enhances i.t. AxE1CAUP/5-FU treatment in hamsters with or without preexisting immunity

Finally, we examined the effects of intermediate-dose CP and preexisting immunity against Ad5 on the efficacy of the i.t. AxE1CAUP/5-FU treatment. This dose of CP efficiently prevented Ad Nab formation (Fig. 5b) and significantly enhanced the antitumor efficacy of both i.t. AxE1CAUP and i.t. AxE1CAUP/5-FU treatments in non-immunized animals (Fig. 5a, p < 0.05 for non-imm/AxE1CAUP/CP vs. non-imm/AxE1CAUP and p < 0.05 for non-imm/AxE1CAUP/CP/5-FU vs. non-imm/AxE1CAUP/5-FU), similarly to the high-dose CP treatment described above (Fig. 2a). Preimmunization by i.m. AxE1CAUP 3 weeks before starting the i.t. AxE1CAUP inoculation significantly (p < 0.01) induced formation of serum Ad Nab (Fig. 5b) and markedly impaired the efficacy of the i.t. AxE1CAUP/5-FU treatment, compared with the efficacy in non-immunized hamsters (Fig. 5a, p < 0.05 for pre-imm/AxE1CAUP/5-FU versus non-imm/AxE1CAUP/5-FU). However, addition of intermediate-dose CP before and during the treatment markedly increased and restored the efficacy of the i.t. AxE1CAUP/5-FU treatment (Fig. 5a, p < 0.01 for pre-imm/AxE1CAUP/CP/5-FU vs. pre-imm/AxE1CAUP/5-FU) to the level of efficacy in the non-imunized animals treated with CP. This was associated with a reduction of the serum Ad Nab titer in the CP-treated animals, regardless of the presence or absence of the preexisting Nab (Fig. 5b).

Figure 5.

Effects of intermediate-dose CP on the efficacy of i.t. AxE1CAUP/5-FU treatment for subcutaneous HaP-T1 tumors in Syrian hamsters with or without preexisting immunity. (a) The effects of preexisting immunity and clinically-approved intermediate-dose of CP (37 mg/kg hamster, equivalent to 5 mg/kg twice weekly in humans[42, 43]) on the antitumor efficacy of i.t. AxE1CAUP/5-FU treatment was examined. The growth of subcutaneous HaP-T1 allografts receiving the following eleven different treatments were compared: (i) non-immunized (non-imm)/PBS (+), (ii) pre-immunized (pre-imm)/PBS (X), (iii) non-imm/CP (▪), (iv) non-imm/5-FU (◻), (v) non-imm/AxE1CAUP /5-FU (△), (vi) non-imm/AxE1CAUP/CP (◊), (vii) non-imm/AxE1CAUP/CP/5-FU (○), (viii) non-imm/AxE1CAUP (*), (ix) pre-imm/AxE1CAUP/5-FU (▴), (x) pre-imm/AxE1CAUP/CP (♦), and (xi) pre-imm/AxE1CAUP/CP/5-FU (•) groups. Mean (n = 6) are shown. Preexisting immunity was generated as described in Material and Methods. *p < 0.05, **p < 0.01. (b) Serum Ad Nab titers following i.t. AxE1CAUP treatment with or without preexisting immunity and intermediate-dose CP. Serum samples were obtained on day 30 from three animals from each group and Ad Nab titers were quantitated as described in Material and Methods. Sera that did not show inhibition of Ad transduction at the dilution lower than (1:16) were considered as undetectable (zero). **p < 0.01.

Discussion

We demonstrated here that AxE1CAUP, an E1B-55 kDa defective OAd expressing UPRT, significantly enhanced the sensitivity of HaP-T1 hamster pancreatic cancer cells to 5-FU in vitro (Fig. 1f) and in vivo after its i.t. injection (Fig. 2a), without causing significant changes in the body weight and the histology of liver and lung (Figs. 4a and 4b) in immunocompetent, Ad-permissive Syrian hamsters. Furthermore, simultaneous addition of high-dose CP inhibited the formation of serum Ad Nab (Fig. 3a), enhanced the replication of AxE1CAUP (Fig. 3b) and the expression of its transgene (UPRT) in the tumors (Figs. 3c and 3d), and subsequently induced more potent antitumor effects following 5-FU administration (Fig. 2a). Finally, we demonstrated that addition of intermediate-dose CP (37 mg/kg twice weekly), which is equivalent to a clinically-approved safe regimen in humans (5 mg/kg twice weekly), similarly enhanced the antitumor effects of the i.t. AxE1CAUP/5-FU treatment in hamsters with or without preexisting Nab against the vector (Fig. 5a). To our knowledge, this is the first report demonstrating the efficacy and safety of an OAd armed with a therapeutic gene and its combination with a clinically-approved safe regimen of CP in an immunocompetent, Ad-permissive animal model. These results indicate the efficacy of CP in overcoming the hurdle of antiviral immunity for successful OV-mediated gene therapy and warrant testing of this “oncolytic gene-chemotherapy approach with immunosuppression” in clinical trials aimed at overcoming the resistance of pancreatic cancers and other malignancies to 5-FU.

Our in vitro experiments showed that the AxE1CAUP vector infects, replicates, causes time-dependent increase in UPRT expression (Figs. 1b, 1c and 1d) in HaP-T1 cells, and resulted in significantly (p < 0.01) enhanced sensitivity of the cells to 5-FU (Figs. 1e and 1f). The sensitizing effect was stronger when 5-FU administration was started late (from 96 hr) than when it was started early (from 24 hr) after the vector inoculation (p < 0.05). This confirmed our previous study, which demonstrated that early exposure to 5-FU before sufficient spread of the vector and its transgene within the tumor caused limited efficacy.[9] In the in vivo study, we observed both mRNA and protein expression of UPRT in the HaP-T1 tumors on day 7 after vector inoculation (Figs. 3c and 3d) and the 5-FU administration started from day 7 significantly (p < 0.01) inhibited the tumor growth (Fig. 2a).

Immunosuppression by CP has previously been reported to prolong transgene expression and to allow successful re-administration of the replication-defective Ad for gene therapy in mouse models.[45, 46] Ikeda et al. have then demonstrated the promoting effects of high-dose CP on the propagation and efficacy of oncolytic herpes simplex virus.[31, 32] In recent studies, CP has also provided other OVs, including OAd,[33] reovirus,[34] measles,[35] and vaccinia virus,[36] with increased oncolytic efficacy. Thomas et al.[33] have shown that high-dose CP inhibits the Ad Nab formation and enhances the replication of OAd and its oncolytic effects in Syrian hamsters. Lamfers et al.[47] have found that high-dose CP increases the transgene (luciferase) expression mediated by an OAd. Here, we extended these observations by demonstrating that the simultaneous use of the same high-dose CP regimen as that used by Thomas et al.[33] inhibited the formation of serum Ad Nab (Fig. 3a), enhanced the replication of the AxE1CAUP vector (Fig. 3b) and the expression of its transgene (UPRT) in tumor tissues (Figs. 3c and 3d), and resulted in significantly (p < 0.01) stronger antitumor effects after 5-FU administration (Fig. 2a).

While we were preparing this manuscript, the article by Peng et al.[42] was E-published, demonstrating that clinically approved intermediate-dose CP regimens also successfully inhibit Nab formation against oncolytic measles and vesicular stomatitis virus, although these authors did not examine their effects on the antitumor efficacy of these OVs. Because of the clinical relevance of their findings, we added experiments examining the effects of intermediate-dose CP (37 mg/kg hamster twice weekly, which is about one-third of the high-dose used and is equivalent to the clinically approved dose of 5 mg/kg twice weekly in humans[42, 43]). We found significantly beneficial effects of this dose of CP, similarly to those of the high-dose CP, on the efficacy of i.t. AxE1CAUP and its combination with 5-FU (Fig. 5a).

More than 80% of adults have been naturally infected with and have Nab for Ad5.[48] The preexisting immunity may significantly reduce the efficacy of OAd. Indeed, the antitumor effects of the i.t. AxE1CAUP/5-FU treatment were significantly (p < 0.05) reduced in hamsters with preexisting Ad Nab induced by previous injection of the vector (Fig. 5a). However, despite the presence of Ad Nab prior to the treatment, the combined use of CP successfully decreased the Ad Nab titer and helped maintain the efficacy of the i.t. AxE1CAUP/5-FU treatment (Figs. 5a and 5b). On the other hand, the only other study that examined the effects of preexisting immunity on an OAd efficacy in immunocompetent Syrian hamsters by Dhar et al.[26] has found no significant effect of preexisting immunity on the efficacy of repetitive (six times) i.t. treatment with wild-type OAd, INGN007 (VRX-007), for syngenic HaK kidney cancer. They also reported that CP administration did not enhance the efficacy of their OAd in the preimmunized hamsters bearing high Ad Nab,[26] although CP did enhance the efficacy in their nonimmunized hamsters.[26, 33] The different results from ours might have arisen from the different OAd used: the VRX-007 used by Dhar et al.[26] was identical to wild-type Ad5 with the exception that it overexpresses the Ad death protein, and thus it replicates and spreads into normal tissues, potentially causing stronger host immune responses than did our AxE1CAUP, which lacks the E1B-55 kDa gene and replicates only in tumor cells. In addition, the antiviral immunity to our AxE1CAUP might have been more restricted because of elimination of the vector by the 5-FU administration as we described previously.[9] Alternatively, the different dose of the OAd used may explain the different results. Since we expected potent cytotoxic effects from the transgene, UPRT, in our OAd, we used only 5% and less than 2% cumulative vector doses of the cumulative doses used by Dhar et al.[26] for preimmunization and treatment, respectively. This indicates that preexisting immunity is a hurdle for lower-dose OAd treatment as ours without using CP, whereas more repetitive, higher-dose OAd treatment, as performed by Dhar et al.,[26] can overcome the preexisting immunity but then the high viral load treatment itself rapidly induces potent antiviral immunity. The relationship between preexisting immunity, therapeutic viral load, and its efficacy requires further in depth study, but may explain why “viral load threshold”[1] was recently found for successful gene delivery to tumors by intravenous (i.v.) injection of oncolytic vaccinia virus, JX-594, in a phase I clinical trial.[49]

In any case, more important than these difference is the fact that Dhar et al. have also shown the benefit of combining high-dose CP for preventing Ad Nab formation and enhancing oncolytic efficacy in their nonimmunized animals,[26, 33] similarly to what was demonstrated here (Figs. 2a and 3a). These results strongly support the contention that the use of CP allows enhanced and repeated OAd therapy. Moreover, in the present study, we made three additional important observations. First, CP not only promoted the oncolytic effects of the Ad vector but also enhanced transgene (UPRT) expression and resulted in stronger combined antitumor efficacy together with 5-FU (Figs. 2a, 3c and 3d). Second, the clinically-approved and safe dose of CP also showed the enhanced antitumor effects. Third, the effect was maintained even in the presence of preexisting immunity to the vector (Fig. 5a), although this third point may depend on the dose, regimen and construction of the virus as discussed above.

Arming the OV with a suicide gene and combining with CP will both enable dose reduction of the OV and inhibition of antiviral immunity, which may allow more repeated treatments. Repeated treatments will be especially beneficial for the treatment using OVs armed with suicide genes, since we and others[7, 9] have previously shown that prodrug administration kills not only the tumor cells but also the OVs themselves because of the potent toxicity of the suicide gene products. Future studies need to confirm the merit of repeating i.t. AxE1CAUP/5-FU treatment by using CP to achieve further enhanced efficacy.

The i.t. AxE1CAUP treatment caused significant (p < 0.01), although mild, growth inhibition of the remote, noninoculated subcutaneous tumors compared with the control tumors (Fig. 2b), suggesting the induction of systemic antitumor immunity. Recently, the ability of various OVs to potentiate the antitumor immune response has received wide attention.[21-23] However, the combined use of CP abolished this beneficial effect of the i.t. AxE1CAUP on the remote tumors (Fig. 2b), suggesting inhibition of the antitumor immunity. Although not addressed in previous studies, this indicates a clear dilemma between antiviral and antitumor immunity:[21-23] although it would be ideal if antitumor immunity could be enhanced, while suppressing antiviral immunity, there is currently no proven way to achieve both of these aims simultaneously.[21-23] If our goal is to induce antitumor immunity, arming the OVs with cytokine genes, rather than suicide genes, will make much more sense. On the other hand, recent recognition of tumor-induced immunosuppression as the major barrier to successful cancer immunotherapy[50] confirmed the efficacy of debulking tumor mass (by surgery, radiotherapy, etc.) before immunotherapy. A high- or intermediate- dose of CP inhibits host immunity,[31-36, 42] whereas a low-dose of CP has been shown to promote antitumor immunity.[28, 30] These considerations lead us to propose the following realistic potential solution for the dilemma. As we suggest in this study, the combination of CP and an OV armed with a suicide gene will serve to reduce the tumor mass via enhanced and repetitive direct tumor-cell killing. Then, subsequent treatment with an OV armed with cytokine genes after restoring the immune system by withdrawing or decreasing the dose of CP should greatly enhance the antitumor immunity. This combined sequential OVs strategy of an immunosuppressive phase using repeated oncolytic suicide-gene therapy, followed by an immunostimulatory phase using oncolytic cytokine-gene therapy deserves testing in the future.

Syrian hamsters are considered suitable animals for evaluating the safety of OAds.[19, 20] We did not detect any immunohistochemical presence of Ad nor any histological damage in the liver and lung specimens obtained from the hamsters treated with AxE1CAUP with or without high-dose CP (Fig. 4b), suggesting the safety of these treatments. These data reflect inability of the AxE1CAUP vector, which lacks the gene for E1B-55 kDa, to replicate in normal cells with preserved p53-signals.[9, 10] In contrast, Dhar et al.[27] reported that i.v. injection of their wild-type Ad, VRX-007, although at high-doses, in combination with CP caused lethal dissemination of the virus to normal organs, which could be prevented by preexisting immunity. Qiao et al. [34] also reported that high-dose CP with i.v. reovirus ablated serum Nab, but this was associated with severe toxicities caused by the systemic replication of the virus. These data indicate that immunosuppressive drugs, such as CP, might be better combined with OVs engineered to replicate only in tumor cells than with wild-type viruses for safety concerns.

We have only tested i.t. injection and have not tested the efficacy of i.v. injection of the AxE1CAUP vector. However, Peng et al.[42] have demonstrated that clinically-approved regimens of CP effectively inhibited the formation of Nab against i.v. injected measles and vesicular stomatitis virus. Therefore, whether clinically-approved regimens of CP can also enhance the efficacy of i.v. injected OVs armed with suicide genes for treating metastatic tumors deserves further studies.

Acknowledgements

This research was supported by Grant-in-Aid for Scientific Research (to M.A.) (C; No. 22590755) and for Young Scientists (to K.F.) (B; No. 21790648) from the Ministry of Education, Culture, Sports, Science and Technology of Japan and by grants from the National Science Council (NSC-100-2320-B-037-020; NSC-101-2320-B-037-047-MY3; NSC-101-2314-B-037-004-My2) and the National Heath Research Institutes (NHRI-EX101-10109BI) of Taiwan, and Kaohsiung Medical University (KMUER-006; KMU-EM-99-3) to K.K.Y.

Ancillary